[
https://issues.apache.org/jira/browse/FLINK-26830?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17511286#comment-17511286
]
Gyula Fora commented on FLINK-26830:
------------------------------------
cc [~wangyang0918] , I hit this while playing around with your PR, I have a fix
that will greately improve this, opening the PR soon
> Move jobs to suspended state before upgrading
> ---------------------------------------------
>
> Key: FLINK-26830
> URL: https://issues.apache.org/jira/browse/FLINK-26830
> Project: Flink
> Issue Type: Sub-task
> Components: Kubernetes Operator
> Reporter: Gyula Fora
> Assignee: Gyula Fora
> Priority: Major
>
> We should not upgrade jobs in one step as that might cause us to lose track
> of what part of the upgrade has succeeded and not.
> For example when upgrading with savepoint strategy we need to record the
> savepoint info in status before trying the deployment because if the
> deployment fails we lose the savepoint info.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
